Functional Safety 2014

– 5

November 2014

Page 3 of 14

Introduction

In the past there has been a preconception that mainly programmable electronic systems will have

systematic failures due to the complexity of the programming system and the unpredictability of

software crashes. Investigation reports over the past few decades have shown that the causes of

some of the major accidents were related to some kind of systematic failures and over 80% of these

accidents are attributable in some degree to human failures, covering both Electro-Mechanical and

Programmable Electronic systems.

This paper looks at why human errors occur and the means of minimising the systematic failures

caused by human errors.

Failure

Failure is defined as the action or state of not functioning; the neglect or omission of the expected or

required action. Failure occurs when a device (or system) does not perform its intended function.

There are two types of failures:-

Physical failures, also known as random hardware failures

Functional failures, also known as systematic failures

Random Hardware Failures

Random hardware failure is the failure of a component, device or system that occurs at a random time.

Random hardware failures are normally well defined and well understood; they can be predicted and

quantified in terms of probability with reasonable accuracy.

The causes of failure are normally a result of material depletion, fatigue or ageing.

Most established manufacturers keep records of their product’s random hardware failures; otherwise

generic failure data from recognisable industrial databases such as NPRD (Nonelectronic Parts

Reliability Data), OREDA® or Exida can be referenced.

Systematic Failures

Systematic failure is a failure that cannot be predicted easily nor quantified statistically. It may occur

while a system is functioning but the system does not perform as intended; or the reason for the

failure may have existed throughout the project phases without being obvious to anyone.

The causes of systematic failures can be due to:

Environmental influences such as flooding, earthquake, storm or electrical interference from

surrounding high voltage equipment;

Human error, such as design faults, inaccurate specification (either safety requirement

specification or design requirement or both), operational errors, ambiguous procedures or

instructions;

Other factors, such as software bugs, software induced failures or incorrect data

communication (e.g. incorrect sequence, data corruption, data loss).